Mobility Disrupted: Geospatial Analysis of Accessibility on the MBTA’s Orange and Red Lines Using Generalized Linear Poisson Regression Author

Abstract

This research analyzes the spatiotemporal relationship between explanatory variables and the frequency of downed elevator/escalator alerts at train stations on the Massachusetts Bay Transportation Authority’s (MBTA) Orange and Red lines. The data is from 2023. The null hypothesis is that there is no spatiotemporal pattern to the frequency/distribution of alerts. Generalized Linear Poisson Regression is used to model relationships across four time periods – the entire 2023 calendar year, Spring (January-April), Summer (May-August), and Fall (September-December). Results indicate that the null hypothesis can be rejected for all variables in Summer and Fall models but cannot be rejected for all variables in entire year and Spring models. Residual and Hot Spot analysis suggest that while technically significant, real-world effects of variable relationships (such as age and/or ridership levels impacting alert frequency) are miniscule.